younesbelkada/transformers

Pull Requests Commits

Add cuda_custom_kernel in DETA (#28989)

SangbumChoi committed 2 years ago

Verified 83e96dc0

younesbelkada committed 2 years ago

1430ca63

Fix static generation when compiling! (#28937)

ArthurZucker committed 2 years ago

Verified f3788b09

[`CLeanup`] Revert SDPA attention changes that got in the static kv cache PR (#29027)

ArthurZucker committed 2 years ago

Verified 609a1767

FIX [`Trainer` / tags]: Fix trainer + tags when users do not pass `"tags"` to `trainer.push_to_hub()` (#29009)

younesbelkada committed 2 years ago

Verified 7a0fccc6

[TPU] Support PyTorch/XLA FSDP via SPMD (#28949)

alanwaketan committed 2 years ago

Verified 5f06053d

Backbone kwargs in config (#28784)

amyeroberts committed 2 years ago

Verified 0199a484

Add tie_weights() to LM heads and set bias in set_output_embeddings() (#28948)

hackyon committed 2 years ago

Verified 725f4ad1

Mask Generation Task Guide (#28897)

merveenoyan committed 2 years ago

Verified 3f4e79d2

Fix flaky test vision encoder-decoder generate (#28923)

zucchini-nlp committed 2 years ago

Verified 354775bc

Introduce AcceleratorConfig dataclass (#28664)

muellerzr committed 2 years ago

Verified 0507e69d

Set the dataset format used by `test_trainer` to float32 (#28920)

ji-huazhong committed 2 years ago

Verified 69ca640d

[`Doc`] Fix docbuilder - make `BackboneMixin` and `BackboneConfigMixin` importable from `utils`. (#29002)

amyeroberts committed 2 years ago

Verified 7252e8d9

AQLM quantizer support (#28928)

Andrei Panferov committed 2 years ago

Verified 1ecf5f7c

Add SiglipForImageClassification and CLIPForImageClassification (#28952)

NielsRogge committed 2 years ago

Verified 63ffd56d

Add `StableLM` (#28810)

jon-tow committed 2 years ago

Verified de6029a0

ENH [`AutoQuantizer`]: enhance trainer + not supported quant methods (#28991)

younesbelkada committed 2 years ago

Verified 164bdef8

ENH: Do not pass warning message in case `quantization_config` is in config but not passed as an arg (#28988)

younesbelkada committed 2 years ago

Verified 1d12b8bc

[`DETR`] Update the processing to adapt masks & bboxes to reflect padding (#28363)

amyeroberts committed 2 years ago

Verified bd4b83e1

Update configuration_llama.py: fixed broken link (#28946)

AdityaKane2001 committed 2 years ago

Verified 3de6a6b4

Static Cache: load models with MQA or GQA (#28975)

gante committed 2 years ago

Verified 3e70a207

Add sudachi_projection option to BertJapaneseTokenizer (#28503)

hiroshi-matsuda-rit committed 2 years ago

Verified da20209d

[`NllbTokenizer`] refactor with added tokens decoder (#27717)

ArthurZucker committed 2 years ago

Verified b4456753

[i18n-de] Translate CONTRIBUTING.md to German (#28954)

khipp committed 2 years ago

Verified d90acc16

[Docs] Add video section (#28958)

NielsRogge committed 2 years ago

Verified 78ba9f46

[Docs] Add language identifiers to fenced code blocks (#28955)

khipp committed 2 years ago

Verified fe3df9d5

Clean up staging tmp checkpoint directory (#28848)

woshiyyya committed 2 years ago

Verified c617f988

Always initialize tied output_embeddings if it has a bias term (#28947)

hackyon committed 2 years ago

Verified 136cd893

Updated requirements for image-classification samples: datasets>=2.14.0 (#28974)

alekseyfa committed 2 years ago

Verified 792819f6

Tests: tag `test_save_load_fast_init_from_base` as flaky (#28930)

gante committed 2 years ago

Verified e30bbb26

Newer Older